* Move to local/stack allocated event API that doesn't require
dynamic allocation/freeing.
* Disable heap, we no longer use alloc/free unless using LVGL.
* Tons of refactors all over to account for the new event approach.
Added BEHAVIOR_DT_DEFINE() and BEHAVIOR_DT_INST_DEFINE(), which work
exactly like the DEVICE_*_DEFINE() macros, except they also register the
device as a behavior by adding a pointer to it to a memory section.
Added zmk_behavior_get_binding(), which works like device_get_binding()
except that it only searches the devices that have been registered as
behaviors. This ensures that behaviors cannot have name collisions with
other devices defined by the SoC, which will be important when we remove
the label property from behaviors so they are given their node names.
As an added benefit, this is faster since it searches a smaller list.
Some basic benchmark code I wrote indicates it takes 30-70% as long,
depending on where the behavior is in the list and whether the name
string is an exact pointer match.
From now on, behaviors should use BEHAVIOR_*_DEFINe() instead of
DEVICE_*_DEFINE(), and any code that looks up a behavior by name should
use zmk_behavior_get_binding() instead of device_get_binding().
Changed all code (except for layer names) which used the label property
to use DEVICE_DT_NAME() instead, which uses the label if set or falls
back to the full node name. This matches how Zephyr determines the node
names used with device_get_binding() and allows us to start removing the
deprecated label property from things.
Detaching the global-quick-tap functionality from the quick-tap term.
This makes way for two improvements:
1. This functionality can be added to combos under a unified name
'global-quick-tap-ms'.
2. This allows users to set a lower term for the 'global-quick-tap'
(typically ~100ms), and a higher term for the regular
quick-tap (typically ~200ms)
This deprecates the global-quick-tap option, however if it is set, the
quick-tap-ms value will be copied to global-quick-tap-ms.
* Move away from DT_LABEL.
* Move to DEVICE_DT_GET for non-behavior device access.
* Move various drivers to `gpio_spec_dt` and `DT` related macros.
* Remove mcp23017 while at it, since better upstream driver is
available.
* This is a very simple fix to a rather complicated issue. Essentially,
hold-taps will "release" (raise) their captured keys before actually
telling the event manager they have captured a key. This means the event
manager ends up assigning the `last_listener_index` to the hold-tap
subscription rather than the combo. So when the combo calls
`ZMK_EVENT_RELEASE` it raises after the hold-tap instead of after the
combo as the combo code expects.
* The corresponding test (which fails without this change) has also been added.
* An event can be captured and released in the same event handler, before
the last_listener_index would have been updated. This causes some handlers
to be triggered multiple times.
* The solution is to update the last_listener_index before calling the next
event handler, so capturing and releasing within an event handler is harmless.
* Also see discussion at https://github.com/zmkfirmware/zmk/pull/1401
* If our handler dedides our undedided hold-tap,
return early before continuing.
* Fix incorrect pointer logic, resulting in combo
candidate filtering leaving incorrect timeout details.
Co-authored-by: Andrew Rae <ajrae.nv@gmail.com>
Co-authored-by: okke <okke@formsma.nl>
* Add optional `hold-trigger-key-positions` hold-tap configuration
* Leverage configuration for decision making around when to trigger
hold decision in hold-taps.
* Add docs for new configuration.
* Tests for the new config/decision logic.
A hold-tap timer event would be triggered too soon if the hold-tap
was delayed for longer than its tapping-term. This may cause
accidental hold behavior when the correct behavior would be tap.
By queuing the timer event instead of executing it immediately,
other delayed events get a chance to be processed properly.
This is an improvement on retro-tap, solving the 'flashing hold' issue
users people experience.
When the tapping-term expires, the hold key is normally pressed. When
retro-tap is enabled, this is undesirable; only an interrupted hold-tap
should trigger the hold behavior.
This change disables the hold behavior for the 'STATUS_HOLD_TIMER'
state when retro-tap is enabled, and makes sure the
'STATUS_HOLD_INTERRUPT' state will be triggered when appropriate.
* Use lesser-known DT features to skip behaviors not referenced
in the user keymap
* Update the behaviors to skip code if no nodes found.
* Remove some empty config/data structs where unused in
behaviors.
Tap-and-hold a hold-tap to hold the tap behavior so it can repeat.
After a tap, if the same key is pressed within `quick_tap_ms`, the
tap behavior is always picked.
This is useful for things like `&ht LSHFT BACKSPACE` where holding
the backspace is required.
Implements #288.
* Make it easier to use *just* event payloads by defining the data,
and then having event manager macros generate "wrapper structs"
* Improve is_*/cast_* APIs to hide details of full event struct.
* Create `zmk_event_t` typedef to pass to event handlers.
* Bring event names inline w/ consistent `zmk_` prefix.
Aligns *.h and *.c to underscore naming convention.
These were kept (with warnings) for backwards compatibility with external boards/shields:
- kscan-mock.h
- matrix-transform.h
They should be removed in the future.
PR: #523
* Add timestamps to position events and behaviors.
- Take original event timestamps into consideration so nested tap-holds have proper timing.
- Add position and timestamp to keycode state changed event so the one-shot behavior can properly identify other keypresses and timings.
- Add timestamp to position events received from peripheral
* reduce number of arguments to behaviors
